Credit union AI has no shortage of promises. Documented deployments with public, linkable sources are harder to come by, and we found two worth writing up: who, what workflow, what tool, what changed, and the caveat to apply before repeating it. Run any candidate this surfaces through the Three-Signals Test before you commit budget.

1. NASA FCU: fraud and AML review

Institution: $5.7 billion in assets, 238,000 members. Workflow: fraud detection, AML alert review, and SAR filing. Tool: DataVisor, expanded from fraud-only into a unified FRAML platform after three years. Outcome: per the credit union’s award announcement, a 42 percent reduction in AML false positives with full true-positive retention, 41 percent less manual review time, and 90 percent faster SAR filing through AI-assisted narratives. Caveat: the numbers are self-reported in an award context and have not been independently audited.

2. Oregonians CU: consumer loan decisioning

Institution: Oregon community credit union, serving members statewide since 1936. Workflow: consumer credit underwriting and off-hours loan decisioning. Tool: Scienaptic AI’s credit decisioning platform. Outcome: per the deployment announcement, instant decisions on nights and weekends, a wider data range than score-only review, and a lending team refocused on exception cases. It is the third credit union Scienaptic announced in three weeks, following Michigan First and Pioneer Valley. Caveat: this one is newly deployed, so treat it as a workflow pattern to watch rather than proven results.

The pattern

Both deployments hit at least two of the three signals from the Three-Signals Test: work that runs constantly on a schedule, is performed nearly identically across people and locations, and resolves in minutes per item. Fraud review queues and consumer loan decisions are exactly that shape. Neither institution started with a member-facing chatbot or a moonshot; both went straight at an internal queue. More candidates worth screening are in the back-office automation pillar.

Get one substantive analysis and one filtered roundup like this every week. Subscribe to the AiForCU newsletter.